On 22/12/06, Sébastien AVELINE saveline@alinto.net wrote:
Hi centos users,
I've installed a load balancer using lvs (with direct routing). I use LVS with a heartbeat configuration and ldirectord and I don't use persistent connections.My problem is when i am running "ipvsadm -lcn", I can see a lot of connections with the CLOSE (or others states) state going from 00:59 to 00:01, and then going back to 00:59. In other words these connections should be dropped after they timed out but the counter is reseted to 60. I wanted to compare these entries on my real servers with netstat and I can say that these connections are not on my real servers and they should be dropped from ip_vs_conn entries. My connection table is growing and I'm wondering if this connections table will not be too huge after a long time.
I use the latest kernel :2.6.9-42.0.3.ELsmp My ipvsadm version is : ipvsadm v1.24 2003/06/07 (compiled with popt and IPVS v1.2.0)
I tried to see with redhat bugzilla but I am still waiting for an answer. I wanted to know if there is a solution and if it will be scheduled for the next kernel (I think that others users camed across this problem). Thanks you for your answers.
Where did you get your ldirector/ipvsadm from? I have 2 sets of 2 load balancers running without any problems at all. They're all
CentOS 4.4 Kernel 2.6.9-42.0.2.EL heartbeat-1.2.3.cvs.20050927-1.centos4 heartbeat-ldirectord-1.2.3.cvs.20050927-1.centos4 ipvsadm-1.24-6
Those heartbeat and ipvsadm RPMs are from the CentOS Extras repositories and have given us next to no bother at all.
http://mirror.centos.org/centos/4/extras/i386/RPMS/
Will.
Will McDonald a écrit :
On 22/12/06, Sébastien AVELINE saveline@alinto.net wrote:
Hi centos users,
I've installed a load balancer using lvs (with direct routing). I use LVS with a heartbeat configuration and ldirectord and I don't use persistent connections.My problem is when i am running "ipvsadm -lcn", I can see a lot of connections with the CLOSE (or others states) state going from 00:59 to 00:01, and then going back to 00:59. In other words these connections should be dropped after they timed out but the counter is reseted to 60. I wanted to compare these entries on my real servers with netstat and I can say that these connections are not on my real servers and they should be dropped from ip_vs_conn entries. My connection table is growing and I'm wondering if this connections table will not be too huge after a long time.
I use the latest kernel :2.6.9-42.0.3.ELsmp My ipvsadm version is : ipvsadm v1.24 2003/06/07 (compiled with popt and IPVS v1.2.0)
I tried to see with redhat bugzilla but I am still waiting for an answer. I wanted to know if there is a solution and if it will be scheduled for the next kernel (I think that others users camed across this problem). Thanks you for your answers.
Where did you get your ldirector/ipvsadm from? I have 2 sets of 2 load balancers running without any problems at all. They're all
CentOS 4.4 Kernel 2.6.9-42.0.2.EL heartbeat-1.2.3.cvs.20050927-1.centos4 heartbeat-ldirectord-1.2.3.cvs.20050927-1.centos4 ipvsadm-1.24-6
Those heartbeat and ipvsadm RPMs are from the CentOS Extras repositories and have given us next to no bother at all.
http://mirror.centos.org/centos/4/extras/i386/RPMS/
Will. _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
We have the same packages, I'm using CentOS 4.4 kernel 2.6.9-42.0.3.ELsmp heartbeat-1.2.3.cvs.20050927 heartbeat-ldirectord-1.2.3.cvs.20050927 ipvsadm 1.24-6 And I took them from extra repository too.
I'm wondering if you are using Direct Routing like me with your set of loadbalancers. I think the problem come from direct routing because loadbalancer can't see all the trafic. But the only thing I am sure is that the counters are reseted after timeout and my connections are not dropped.
On 22/12/06, Sébastien AVELINE saveline@alinto.net wrote:
Will McDonald a écrit :
On 22/12/06, Sébastien AVELINE saveline@alinto.net wrote:
Hi centos users,
I've installed a load balancer using lvs (with direct routing). I use LVS with a heartbeat configuration and ldirectord and I don't use persistent connections.My problem is when i am running "ipvsadm -lcn", I can see a lot of connections with the CLOSE (or others states) state going from 00:59 to 00:01, and then going back to 00:59. In other words these connections should be dropped after they timed out but the counter is reseted to 60. I wanted to compare these entries on my real servers with netstat and I can say that these connections are not on my real servers and they should be dropped from ip_vs_conn entries. My connection table is growing and I'm wondering if this connections table will not be too huge after a long time.
I use the latest kernel :2.6.9-42.0.3.ELsmp My ipvsadm version is : ipvsadm v1.24 2003/06/07 (compiled with popt and IPVS v1.2.0)
I tried to see with redhat bugzilla but I am still waiting for an answer. I wanted to know if there is a solution and if it will be scheduled for the next kernel (I think that others users camed across this problem). Thanks you for your answers.
Where did you get your ldirector/ipvsadm from? I have 2 sets of 2 load balancers running without any problems at all. They're all
CentOS 4.4 Kernel 2.6.9-42.0.2.EL heartbeat-1.2.3.cvs.20050927-1.centos4 heartbeat-ldirectord-1.2.3.cvs.20050927-1.centos4 ipvsadm-1.24-6
Those heartbeat and ipvsadm RPMs are from the CentOS Extras repositories and have given us next to no bother at all.
http://mirror.centos.org/centos/4/extras/i386/RPMS/
Will. _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
We have the same packages, I'm using CentOS 4.4 kernel 2.6.9-42.0.3.ELsmp heartbeat-1.2.3.cvs.20050927 heartbeat-ldirectord-1.2.3.cvs.20050927 ipvsadm 1.24-6 And I took them from extra repository too.
I'm wondering if you are using Direct Routing like me with your set of loadbalancers. I think the problem come from direct routing because loadbalancer can't see all the trafic. But the only thing I am sure is that the counters are reseted after timeout and my connections are not dropped.
I'm afraid not, all our systems are setup using masquerading NAT. That was going to be my next question if it turned out (as it did) that you're using kosher RPMs. :)
I've only ever had LVSes configured with NAT so have no experience and only a vague memory of how DR works from the docs. If no one here can point you in the right direction there's a dedicated LVS list too which might be worth searching the archives of and then questioning if you can't find an answer.
http://archive.linuxvirtualserver.org/html/lvs-users/
Will.
On 22/12/06, Will McDonald wmcdonald@gmail.com wrote:
On 22/12/06, Sébastien AVELINE saveline@alinto.net wrote:
Will McDonald a écrit :
On 22/12/06, Sébastien AVELINE saveline@alinto.net wrote:
Hi centos users,
I've installed a load balancer using lvs (with direct routing). I use LVS with a heartbeat configuration and ldirectord and I don't use persistent connections.My problem is when i am running "ipvsadm -lcn", I can see a lot of connections with the CLOSE (or others states) state going from 00:59 to 00:01, and then going back to 00:59. In other words these connections should be dropped after they timed out but the counter is reseted to 60. I wanted to compare these entries on my real servers with netstat and I can say that these connections are not on my real servers and they should be dropped from ip_vs_conn entries. My connection table is growing and I'm wondering if this connections table will not be too huge after a long time.
I'm afraid not, all our systems are setup using masquerading NAT. That was going to be my next question if it turned out (as it did) that you're using kosher RPMs. :)
I've only ever had LVSes configured with NAT so have no experience and only a vague memory of how DR works from the docs. If no one here can point you in the right direction there's a dedicated LVS list too which might be worth searching the archives of and then questioning if you can't find an answer.
And, to follow up my own post, I've just had a quick look through my LVS list mails and spotted:
http://archive.linuxvirtualserver.org/html/lvs-users/2006-11/msg00200.html
There's no follow up from the OP which may or may not be a good thing. :)
It appears a kernel upgrade to 2.6.18 (!) may help. You might want to fire off a mail to the OP and ask if he had any success. I have the un-obfuscated address in my mailstore, I'll send it offlist.
Will.
Will McDonald a écrit :
On 22/12/06, Will McDonald wmcdonald@gmail.com wrote:
On 22/12/06, Sébastien AVELINE saveline@alinto.net wrote:
Will McDonald a écrit :
On 22/12/06, Sébastien AVELINE saveline@alinto.net wrote:
Hi centos users,
I've installed a load balancer using lvs (with direct routing).
I use
LVS with a heartbeat configuration and ldirectord and I don't use persistent connections.My problem is when i am running "ipvsadm -lcn", I can see a lot of connections with the CLOSE (or others states) state going from 00:59 to 00:01, and then going back to 00:59.
In other
words these connections should be dropped after they timed out
but the
counter is reseted to 60. I wanted to compare these entries on my real
servers with
netstat and I can say that these connections are not on my real servers and they should be dropped from ip_vs_conn entries. My connection table is growing and I'm wondering if this connections table will not be too huge after a
long
time.
I'm afraid not, all our systems are setup using masquerading NAT. That was going to be my next question if it turned out (as it did) that you're using kosher RPMs. :)
I've only ever had LVSes configured with NAT so have no experience and only a vague memory of how DR works from the docs. If no one here can point you in the right direction there's a dedicated LVS list too which might be worth searching the archives of and then questioning if you can't find an answer.
And, to follow up my own post, I've just had a quick look through my LVS list mails and spotted:
http://archive.linuxvirtualserver.org/html/lvs-users/2006-11/msg00200.html
There's no follow up from the OP which may or may not be a good thing. :)
It appears a kernel upgrade to 2.6.18 (!) may help. You might want to fire off a mail to the OP and ask if he had any success. I have the un-obfuscated address in my mailstore, I'll send it offlist.
Will. _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos
Thanks for your help, but I will wait for an official patch from centos, I hope that I will be heard.
Sebastien
On Tue, 2006-12-26 at 10:25 +0100, Sébastien AVELINE wrote:
<SNIP>
And, to follow up my own post, I've just had a quick look through my LVS list mails and spotted:
http://archive.linuxvirtualserver.org/html/lvs-users/2006-11/msg00200.html
There's no follow up from the OP which may or may not be a good thing. :)
It appears a kernel upgrade to 2.6.18 (!) may help. You might want to fire off a mail to the OP and ask if he had any success. I have the un-obfuscated address in my mailstore, I'll send it offlist.
Will.
Thanks for your help, but I will wait for an official patch from centos, I hope that I will be heard.
Sebastien
Please follow the above thread and let me know if / when they find the problem.
We will not be upgrading "officially" the ipvs RPM above the one that is in the RHCS / RHGFS RPMS ... but we can put something in the testing repo if ipvs turns out to be the issue.
We can also feedback via the RHEL bugzilla to get any patches incorporated (as required) to the ipvs RPM upstream.
Thanks, Johnny Hughes
-
Johnny Hughes -
Sébastien AVELINE -
Will McDonald